Latest-generation CCTV cameras are bringing intelligent observation to some of the world’s most scrutinised cities. E&T zooms in on the innovations.
Nobody knows for sure how many CCTV security cameras there are installed in the UK alone (between three million and six million is a best estimate) watching streets, airports, hospitals, shops, rail platforms, car parks, entrances and exits, office corridors and housing estates, and all manner of other urban nooks and crannies.
Around 90 per cent of those run on analogue technology, and form our picture of the typical swivelling Big Brother surveillance 'eye' but increasing availability of lower-cost advanced digital network cameras able to capture and store high-definition (HD) video images, and send them over standard IP (Internet protocol) networks, looks set to make a significant impact on that ratio, as well as revolutionise the role CCTV plays in crime detection.
Analogue CCTV cameras have reached the peak of their technological prowess as far as security surveillance is concerned. Analyst Frost & Sullivan expects sales of IP video surveillance cameras to exceed those of analogue across Europe by 2012 (partly because IP cameras tend to cost more), with sales of analogue equipment declining more slowly than anticipated only because the market itself is expanding – $765m (£490m) in 2009 to $955m (£610m) by 2015.
The new-generation cameras bring many advances on analogue. Foremost is higher picture resolution, which provides digital surveillance cameras with sharper image capture to make it easier for operators to clearly identify incidents captured on camera, and put an end to the blurred, pixellated analogue CCTV images of suspects we used to see on TV crime-watch programmes. Better definition also means that cameras can be sited at a greater distance from the area being surveilled, and are therefore less noticeable.
'If it is easy to put a camera up, but when you come to view the images you cannot see who people are, or recognise number plates and so on, it is absolutely useless as an investment,' says Phil Doyle, UK and Ireland managing director at network camera specialist, Axis Communications.
'Some 90 per cent of UK CCTV installations still record images at no more than 0.1 megapixels, and resolution is key in avoiding block images,' added Mike Lewis, UK and Eire country manager for digital security firm MOBOTIX.
Pure numbers-based comparisons of analogue versus digital CCTV resolutions are hard to precisely define, because analogue relies on video line standards like PAL (Europe) and NSTC (US) that use different types of pixels suitable for television displays rather than computer monitors, as well as a variety of common international formats (CIFs). The fact that analogue images are comprised of interlaced video, while digital images are captured using progressive scan technology, is another contributory factor. Generally speaking, however, the best video resolution analogue can practicably provide is 768x576 or 720x480 pixels, whereas digital IP cameras tend to offer VGA (640x480), SVGA (800x600), and increasingly quad VGA (or megapixel) resolutions of 1280x960 pixels equivalent to HDTV pictures.
Some digital CCTV cameras can capture still video images at resolutions of up to 5 megapixels already, but at very low frame rates that deliver footage more as a serious of connected pictures rather than full-motion video. Given that digital camera technology is advancing all the time, there seems little to stop those same cameras acquiring enough processing power to deliver even higher digital video resolutions as the technology progresses.
The other advantage of HD video surveillance is that, in some cases, fewer digital cameras can scope the same area, with high-speed optical and digital zoom lens and functions meaning little or no loss of quality when focusing on specific image points.
'With the HD standard you do not have to zoom in as far [because the image is already clearer], but the camera also provides wider coverage and you do not necessarily need the same amount of cameras,' explains Doyle. This is a key factor in promoting transition from analogue to digital.
Data storage concerns
The number of frames per second (fps) at which video images are captured also has a significant effect on overall image quality, and crucially the back-end IT resources needed to process and store that data for however many hours, days, weeks or months, depending on the requirements of the organisation deploying them, which represents a big cost of the overall system.
Briefly put, the better the image quality, the more the data generated. Video footage captured at 30fps will use far more storage than equivalent images, at the same resolution, captured at 15fps for example; but the lower the fps rate, the longer the gaps between frames which can limit the ability to identify moving objects in particular.
While a higher frame rate means more processing power and disk storage space is required, not all digital surveillance operations require high frame rates, and most digital IP cameras offer the ability to experiment with different resolutions and frame rates to match different requirements.
'The human eye is only able to detect the image at around 17fps – the only time you will need more than that is where you need to capture and analyse in detail an incident that happened very fast, such as sleight-of-hand in a casino,' reckons Lewis. 'For retail monitoring, for example, 12fps is more than enough.'
Video signals usually are not transmitted in their raw state: a form of data compression is applied on the fly. The compression format, or Codec, is also key to balancing video quality against data storage capacity, with most digital cameras today supporting common standards such as H.264 and MPEG-4.
The development of Power over Ethernet (PoE) technology also means that wired digital surveillance cameras do not need a separate power source – they can draw up to 13w of power just from a standard Cat5e/6 Ethernet data network cable, which simplifies their installation to a great extent. Wireless cameras still require their own power supply, though the wireless access points which connect them into the wired network can also be powered via PoE.
Having only a single cable connection that handles both IP data and power helps reduce installation costs. The convenience of being able to pipe IP CCTV traffic across an organisation's existing voice/data network is another key advantage of digital video surveillance technology, but only to a point. Depending on the number of cameras transmitting images at any one time, and the resolution and frame rate those images are captured at, that traffic can take up a lot of bandwidth, potentially depriving other business applications of the network capacity they need to function properly.
'When network video was first invented in the mid-1990s, it was intended that it would use the same local area network (LAN),' says Doyle. 'On smaller installations they still run on the same LAN as other corporate applications, but we are seeing lots of projects that involve hundreds or thousands of cameras which use a parallel network to keep management and operations separate.' Again, where the technology brings benefits at one end of the IP network it imposes new complications at the other.
The volume of video footage transmitted over the network can be reduced by using data storage facilities integrated into the camera itself. Most cameras provide the option of sending video images to a centralised storage resource, usually network-attached server or PC hard disks located in the control room, and in some cases dedicated digital video recorders. Others add local data storage facilities, such as integral Flash memory or removable media, which provide a backup copy in the event of server or network failure, or temporary data storage facilities that are automatically deleted after a certain time period, but which can be retrieved and reviewed should an incident occur.
In fact, Doyle says, it is pretty much an even split between digital surveillance cameras that process and store video footage locally, and those which simply relay images back to server hard disks for remote analytics and storage.
Building 'intelligent' processing capabilities into the camera itself provides a number of other advantages, including the ability to integrate motion detection and heat sensors, speakers, and microphones. Other cameras often provide AV ports, which allow similar devices to be plugged in. These sensors can detect movement or changes in ambient conditions that trigger recording according to pre-defined parameters, cutting-down on the volume of footage that needs to be processed and stored.
Armed with powerful CPUs, the cameras can also process captured images themselves, and make a decision about whether to retain that footage or overwrite it depending on the incidents it detects. The camera can either record nothing at all if its sensors decide there is nothing worth recording; or it can record at a higher resolution and frame rate to make sure it captures any specific action in better detail. 'You might put one in the office ceiling that records at a certain resolution and frame rate whenever somebody is in the room, or hook them up to temperature and fan sensors that trigger recordings based on heat or flames,' says Lewis.
An example is the Bosch AutoDome Easy II model: this camera has integrated video analysis software detects loitering, idle objects, object removal and line-crossing, and sends alerts to system administrators.
Axis Communications does not use its own video analytics software, but provides an open application programming interface that allows third part developers to create their own. 'You can have crowd management detection that sends alerts when multiple people are detected, for example, or have software that detects if somebody is standing too close in a queue for the ATM indicating they are trying to copy a pin number,' Doyle points out.
Some cameras can be configured to send out some sort of an alert to a control room or individual, via email or SMS for example, and automatically start streaming images to an off-site location where they cannot be destroyed when pre-defined parameters indicating likely incidents are met. Remote operators can also use the integral speaker to warn off intruders when they are detected by informing them they are being recorded.
Digital CCTV scenarios
The majority of digital IP surveillance devices are used for security monitoring of individuals in shops, city centres, airports, car parks, police stations and other public areas. But as the technology has evolved, so have innovative new usage scenarios, including automatic number plate recognition for congestion charging now used in London and other European cities, till monitoring inside shops to guard against employee theft, and track haulage and construction company yard operations.
Other uses of the technology are based around health and safety initiatives. Sussex utility company Southern Water is using IP CCTV to provide visual verification for its drinking water sampling procedures at up to 100 sites across the south of the UK, for example.
Connected to the company's head office via ADSL or MPLS broadband links, the cameras send alerts when they detect an engineer entering a water sampling kiosk so a remote operator can monitor the process in real time. One of the biggest vertical markets for HD digital surveillance cameras is education – mostly schools and colleges but also universities – partly for anti-theft and anti-vandalism purposes but also to make sure the teachers are behaving themselves.
Cameras are also used to watch how teachers talk to children and vice versa to prevent abuse, as well as pro-active teacher and classroom training, and quality of teaching monitoring. Users can also talk back into the classroom through a microphone and speaker embedded in the camera.
Digital security cameras with local data storage facilities are also finding their way onto different forms of moving transport, such as trains and buses. One Newcastle bus company has mounted a 360° view camera in the centre of the buses roof which constantly monitors the activity of passengers and staff, and stores the data on an integrated SD card. Once back in the depot, the SD card can be removed for analysis or the footage downloaded wirelessly to a company computer for analysis, if required.